You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@httpd.apache.org by Rob Hartill <ro...@imdb.com> on 1996/11/23 15:15:09 UTC
Don't know other ways (fwd)
Problems getting the docs..
----- Forwarded message from smelekov@mephi.ru -----
Date: Sat, 23 Nov 1996 11:54:53 -0300
Message-Id: <96...@eagle.dec-center.mephi.ru>
To: apache-bugs@mail.apache.org
X-Url: mailto:apache-bugs@mail.apache.org
X-Mailer: Lynx, Version 2-4-2
X-Personal_Name: Serguei Melekhov
From: smelekov@mephi.ru
Subject: Don't know other ways
Hello. Here is my problem. As I discovered there is no information files on your
servers that totally describes apache1.1. I mean regular text file. You have
only .dvi .tex. It is unconvinient. Please if you can help me with that e-mail
me back full info bout apache 1.1 in .txt file. Connection is slow enough
not to let us to ftp something from your server. It will take
lot of time...
Yours,
Serguei Melekhocv
Serguei Melekhov
----- End of forwarded message from smelekov@mephi.ru -----
Re: Don't know other ways (fwd)
Posted by Paul Richards <p....@elsevier.co.uk>.
Brian Behlendorf <br...@organic.com> writes:
> Let's get our terms straight. SGML is a document meta-language. HTML is
> defined in a Document Type Definition, written in SGML. Documents are not "in
> SGML format", they are in a document language described by an SGML DTD. By
> definition, HTML is an "SGML format". So, I'm confused - are you saying we
> should create our own DTD, and author within that? I don't see significant
> value of that over using a well-defined subset of HTML 2.0 or 3.2 (i.e.
> avoiding <FONT>, etc). If there's a need for richer semantics, we could use
> CLASSes on tags, and then apply style sheets to those classes. The costs in my
> opinion of switching to our own DTD would be in reduced access (those who edit
> and view the source docs must use SGML tools), having to decide upon a new DTD
> (a political/engineering process), and we'd still have to define how the new
> DTD maps to PS, LaTex, etc. All that work has been done for us with HTML.
I know that technically HTML is in "SGML format" but it's a really bad
DTD that even when strictly adhered to doesn't sit well with the
concepts of SGML in that a lot of the tags are to do with presentation
and not structure.
In practice, if we used HTML as our "source" DTD then we'd have very
boring web pages and the breadth of structure available for our
docuements would be restricted.
I'm not thinking in terms of just producing hard-copy of the web
pages, I'm thinking in terms of producing the manual as a book or as
info-pages etc. You want to be able to tag something as a chapter
heading or a section heading, rather than <H1> which says nothing
about what the tagged text actually is. Then your mapping file can
decide how to display a chapter heading in whatever the output format
is. We could use <H1> to mean "chapter heading" but we'd have to
overload the tags since html doesn't have a particularly rich set of
structure tags since as I said above, its tags are overly geared
towards presentation.
We wouldn't need to create our own DTD, there are plenty out there
already. FreeBSD is moving from a linux DTD (which is basically Latex
markup as SGML tags) to docbook. The tools for conversion and all the
mappings would be there to just pick up. I could find out how the
XFree86 team do their docs since that's more comparable to FreeBSD and
they use SGML as well.
I agree that the source would be less accessible in that people here
will be familiar with html but my experience from FreeBSD is that
writing docs using some other DTD is not a problem since most
computer folks are quite comfortable with the concept of tagging text
(either from html or latex) and using different tags hasn't been a
problem particularly if they're tags such as <chapter> etc.
--
Paul Richards. Originative Solutions Ltd. (Netcraft Ltd. contractor)
Elsevier Science TIS online journal project.
Email: p.richards@elsevier.co.uk
Phone: 0370 462071 (Mobile), +44 (0)1865 843155
Re: Don't know other ways (fwd)
Posted by Ed Korthof <ed...@organic.com>.
I agree with Brian -- we don't want to do SGML. I've worked some with SGML --
it's not easy to use, tools are hard to find, hard to use, and/or expensive,
and it seems unlikely that there are enough people familiar with it who are
documenting Apache for a switchover to be possible w/o lots of time and effort.
*If* we really want to maintain extra information, Brian's suggestion seems
reasonable (though I'm not very familiar with CLASSes or stylesheets, so long
as the tags contain all the structure, they'd be fine; tagged-text processing
is considerably easier than full SGML authoring; the only disadvantage is a
lack of enforcement of rules -- which we should do ourselves anyway). With any
set of tags which indicates the relevant structure, we'd easily be able to
extract all the information SGML would allow.
Ed Korthof
ed@organic.com
On Nov 25, 10:25am, Brian Behlendorf wrote:
> Let's get our terms straight. SGML is a document meta-language. HTML is
> defined in a Document Type Definition, written in SGML. Documents are not
"in
> SGML format", they are in a document language described by an SGML DTD. By
> definition, HTML is an "SGML format". So, I'm confused - are you saying we
> should create our own DTD, and author within that? I don't see significant
> value of that over using a well-defined subset of HTML 2.0 or 3.2 (i.e.
> avoiding <FONT>, etc). If there's a need for richer semantics, we could use
> CLASSes on tags, and then apply style sheets to those classes. The costs in
my
> opinion of switching to our own DTD would be in reduced access (those who
edit
> and view the source docs must use SGML tools), having to decide upon a new
DTD
> (a political/engineering process), and we'd still have to define how the new
> DTD maps to PS, LaTex, etc. All that work has been done for us with HTML.
>
> Brian
Re: Don't know other ways (fwd)
Posted by Brian Behlendorf <br...@organic.com>.
On Mon, 25 Nov 1996 rasmus@mail1.bellglobal.com wrote:
> > Clean HTML can be the source for other document formats, in my opinion. We
> > just have to avoid using a lot of fancy unportable tags.
>
> But a clean HTML document is SGML. It would certainly be more convenient to
> have the docs in SGML format.
Let's get our terms straight. SGML is a document meta-language. HTML is
defined in a Document Type Definition, written in SGML. Documents are not "in
SGML format", they are in a document language described by an SGML DTD. By
definition, HTML is an "SGML format". So, I'm confused - are you saying we
should create our own DTD, and author within that? I don't see significant
value of that over using a well-defined subset of HTML 2.0 or 3.2 (i.e.
avoiding <FONT>, etc). If there's a need for richer semantics, we could use
CLASSes on tags, and then apply style sheets to those classes. The costs in my
opinion of switching to our own DTD would be in reduced access (those who edit
and view the source docs must use SGML tools), having to decide upon a new DTD
(a political/engineering process), and we'd still have to define how the new
DTD maps to PS, LaTex, etc. All that work has been done for us with HTML.
Brian
--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@organic.com www.apache.org hyperreal.com http://www.organic.com/JOBS
Re: Don't know other ways (fwd)
Posted by ra...@mail1.bellglobal.com.
> Clean HTML can be the source for other document formats, in my opinion. We
> just have to avoid using a lot of fancy unportable tags.
But a clean HTML document is SGML. It would certainly be more convenient to
have the docs in SGML format.
-Rasmus
Re: Don't know other ways (fwd)
Posted by Brian Behlendorf <br...@organic.com>.
On 25 Nov 1996, Paul Richards wrote:
> Rob Hartill <ro...@imdb.com> forwards:
>
> > servers that totally describes apache1.1. I mean regular text file. You have
> > only .dvi .tex. It is unconvinient. Please if you can help me with that e-mail
> > me back full info bout apache 1.1 in .txt file. Connection is slow enough
> > not to let us to ftp something from your server. It will take
> > lot of time...
>
> We should consider switching to SGML source for 2.0. This is what most
> of the other free software projects have done, we can generate html,
> ps, ascii, whatever from it then.
Clean HTML can be the source for other document formats, in my opinion. We
just have to avoid using a lot of fancy unportable tags.
Brian
--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@organic.com www.apache.org hyperreal.com http://www.organic.com/JOBS
Re: Don't know other ways (fwd)
Posted by Paul Richards <p....@elsevier.co.uk>.
Rob Hartill <ro...@imdb.com> writes:
> servers that totally describes apache1.1. I mean regular text file. You have
> only .dvi .tex. It is unconvinient. Please if you can help me with that e-mail
> me back full info bout apache 1.1 in .txt file. Connection is slow enough
> not to let us to ftp something from your server. It will take
> lot of time...
We should consider switching to SGML source for 2.0. This is what most
of the other free software projects have done, we can generate html,
ps, ascii, whatever from it then.
--
Paul Richards. Originative Solutions Ltd. (Netcraft Ltd. contractor)
Elsevier Science TIS online journal project.
Email: p.richards@elsevier.co.uk
Phone: 0370 462071 (Mobile), +44 (0)1865 843155